DOCSLIB.ORG

Virtual Machine Workloads: the Case for New Benchmarks For

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Virtual Machine Workloads: the Case for New Benchmarks For Virtual Machine Workloads: The Case for New Benchmarks for NAS Vasily Tarasov1, Dean Hildebrand2, Geoff Kuenning3, Erez Zadok1 1Stony Brook University, 2IBM Research—Almaden, 3Harvey Mudd College Abstract images being stored on NAS servers. Encapsulating file systems in virtual disk image files simplifies the imple- Network Attached Storage (NAS) and Virtual Ma- mentation of features such as migration, cloning, and chines (VMs) are widely used in data centers thanks snapshotting, since they naturally map to existing NAS to their manageability, scalability, and ability to con- functions. In addition, non-virtualized hosts can co-exist solidate resources. But the shift from physical to vir- peacefully with virtualized ones that use the same NAS tual clients drastically changes the I/O workloads seen interface, which permits a gradual migration of services on NAS servers, due to guest file system encapsula- from physical to virtual machines. tion in virtual disk images and the multiplexing of re- quest streams from different VMs. Unfortunately, cur- Storage performance plays a crucial role when ad- rent NAS workload generators and benchmarks produce ministrators select the best NAS for their environment. workloads typical to physical machines. One traditional way to evaluate NAS performance is to This paper makes two contributions. First, we studied run a file system benchmark, such as SPECsfs2008 [38]. the extent to which virtualization is changing existing Vendors periodically submit the results of SPECsfs2008 NAS workloads. We observed significant changes, in- to SPEC; the most recent submission was in Novem- cluding the disappearance of file system meta-data op- ber 2012. Because widely publicized benchmarks such erations at the NAS layer, changed I/O sizes, and in- as SPECsfs2008 figure so prominently in configuration creased randomness. Second, we created a set of ver- and purchase decisions, it is essential to ensure that the satile NAS benchmarks to synthesize virtualized work- workloads they generate represent what is observed in loads. This allows us to generate accurate virtualized real-world data centers. workloads without the effort and limitations associated This paper makes two contributions: an analysis of with setting up a full virtualized environment. Our ex- changing virtualized NAS workloads, and the design and periments demonstrate that the relative error of our virtu- implementation of a system to generate realistic virtu- alized benchmarks, evaluated across 11 parameters, av- alized NAS workloads. We first demonstrate that the erages less than 10%. workloads generated by many current file system bench- marks do not represent the actual workloads produced by 1 Introduction VMs. This in turn leads to a situation where the perfor- By the end of 2012 almost half of all applications run- mance results of a benchmark deviate significantly from ning on x86 servers will be virtualized; in 2014 this the performance observed in real-world deployments. number is projected to be close to 70% [8,9]. Virtualiza- Although benchmarks are never perfect models of real tion, if applied properly, can significantly improve sys- workloads, the introduction of VMs has exacerbated the tem utilization, reduce management costs, and increase problem significantly. Consider just one example, the system reliability and scalability. With all the benefits percentage of data and meta-data operations generated of virtualization, managing the growth and scalability of by physical and virtualized clients. Table 1 presents the storage is emerging as a major challenge. results for the SPECsfs2008 and Filebench web-server In recent years, growth in network-based storage has benchmarks that attempt to provide a “realistic” mix of outpaced that of direct-attached disks; by 2014 more meta-data and data operations. We see that meta-data than 90% of enterprise storage capacity is expected to be procedures, which dominated in physical workloads, are served by Network Attached Storage (NAS) and Storage almost non-existent when VMs are utilized. The reason Area Networks (SAN) [50]. Network-based storage can is that VMs store their guest file system inside large disk improve availability and scalability by providing shared image files. Consequently, all meta-data operations (and access to large amounts of data. Within the network- NFS Physical clients Virtualized based storage market, NAS capacity is predicted to in- procedures (SPECsfs2008/Filebench) clients crease at an annual growth rate of 60%, as compared to Data 28% / 36% 99% only 22% for SAN [43]. This faster NAS growth is ex- Meta-data 72% / 64% <1% plained in part by its lower cost and its convenient file Table 1: The striking differences between virtualized and system interface, which is richer, easier to manage, and physical workloads for two benchmarks: SPECsfs2008 and more flexible than the block-level SAN interface. Filebench (Web-server profile). Data operations include READ The rapid expansion of virtualization and NAS has and WRITE. All other operations (e.g., CREATE, GETATTR, lead to explosive growth in the number of virtual disk READDIR) are characterized as meta-data. 1 USENIX Association 11th USENIX Conference on File and Storage Technologies (FAST ’13) 307 indeed all data operations) from the applications are con- VM 2 Files on NAS verted into simple reads and writes to the image file. Guest OS Meta-data-to-data conversion is just one example of 1a Disk image file on DAS the way workloads shift when virtual machines are in- troduced. In this paper we examine, by collecting Local On−disk Network−based 1b Disk image file on SAN File System File System and analyzing a set of I/O traces generated by current Driver for benchmarks, how NAS workloads change when used I/O Controller in virtualized environments. We then leverage multi- 1c Disk image file on NAS Case analyzed in this paper dimensional trace analysis techniques to convert these Emulated traces to benchmarks [13, 40]. Our new virtual bench- I/O Controller 1d Pass−through to DAS or SAN marks are flexible and configurable, and support single- and multi-VM workloads. With multi-VM workloads, Emulated Disk the emulated VMs can all run the same or different ap- plication workloads (a common consequence of resource Hypervisor consolidation). Further, users do not need to go through Figure 1: VM data-access methods. Cases 1a–1d correspond a complex deployment process, such as hypervisor setup to the emulated-block-device architecture. Case 2 corresponds and per-VM OS and application installation, but can in- to the use of guest network file system clients. stead just run our benchmarks. This is useful because administrators typically do not have access to the pro- 2.1.1 Emulated Block Devices duction environment when evaluating new or existing NAS servers for prospective virtualized clients. Finally, Figure 1 shows several options for implementing the some benchmarks such as SPECsfs cannot be usefully back end of an emulated block device: run inside a VM because they do not support file-level 1a. A file located on a local file system that is deployed interfaces and will continue to generate a physical work- on Direct Attached Storage (DAS). This approach is load to the NAS server; this means that new benchmarks used, for example, by home and office installations of can be the only viable evaluation option. Our bench- VMware Workstation [39] or Oracle VirtualBox [44]. marks are capable of simulating a high load (i.e., many Such systems often keep their disk images on local file VMs) using only modest resources. Our experiments systems (e.g., Ext3, NTFS). Although this architecture demonstrate that the accuracy of our benchmarks re- works for small deployments, it is rarely used in large mains within 10% across 11 important parameters. enterprises where scalability, manageability, and high availability are critical. 2 Background 1b. A disk image file is stored on a (possibly clus- In this section, we present several common data ac- tered) file system deployed over a Storage Area Network cess methods for virtualized applications, describe in (SAN) (e.g., VMware’s VMFS file system [46]). A SAN depth the changes in the virtualized NAS I/O stack (VM- offers low-latency shared access to the available block NAS), and then explain the challenges in benchmarking devices, which allows high-performance clustered file NAS systems in virtualized environments. systems to be deployed on top of the SAN. This archi- tecture simplifies VM migration and offers higher scal- 2.1 Data Access Options for VMs ability than DAS, but SAN hardware is more expensive Many applications are designed to access data using a and complex to administer. conventional POSIX file system interface. The methods 1c. A disk image file stored on Network Attached Stor- that are currently used to provide this type of access in a age (NAS). In this architecture, which we call VM-NAS, VM can be classified into two categories: (1) emulated the host’s hypervisor passes I/O requests from the virtual block devices (typically managed in the guest by a local machine to an NFS or SMB client, which in turn then ac- file system); and (2) guest network file system clients. cesses a disk image file stored on an external file server. Figure 1 illustrates both approaches. With an emu- The hypervisor is completely unaware of the storage ar- lated block device, the hypervisor emulates an I/O con- chitecture behind the NAS interface. NAS provides the troller with a connected disk drive. Emulation is com- scalability, reliability, and data mobility needed for ef- pletely transparent to the guest OS, and the virtual I/O ficient VM management. Typically, NAS solutions are controller and disk drives appear as physical devices to cheaper than SANs due to their use of IP networks, and the OS. The guest OS typically formats the disk drive are simpler to configure and manage.
Load more

Recommended publications
  • The Nutanix Design Guide
    The FIRST EDITION Nutanix Design Guide Edited by AngeloLuciani,VCP By Nutanix, RenévandenBedem, incollaborationwith NPX,VCDX4,DECM-EA RoundTowerTechnologies Table of Contents 1 Foreword 3 2 Contributors 5 3 Acknowledgements 7 4 Using This Book 7 5 Introduction To Nutanix 9 6 Why Nutanix? 17 7 The Nutanix Eco-System 37 8 Certification & Training 43 9 Design Methodology 49 & The NPX Program 10 Channel Charter 53 11 Mission-Critical Applications 57 12 SAP On Nutanix 69 13 Hardware Platforms 81 14 Sizer & Collector 93 15 IBM Power Systems 97 16 Remote Office, Branch Office 103 17 Xi Frame & EUC 113 Table of Contents 18 Xi IoT 121 19 Xi Leap, Data Protection, 133 DR & Metro Availability 20 Cloud Management 145 & Automation: Calm, Xi Beam & Xi Epoch 21 Era 161 22 Karbon 165 23 Acropolis Security & Flow 169 24 Files 193 25 Volumes 199 26 Buckets 203 27 Prism 209 28 Life Cycle Manager 213 29 AHV 217 30 Move 225 31 X-Ray 229 32 Foundation 241 33 Data Center Facilities 245 34 People & Process 251 35 Risk Management 255 The Nutanix Design Guide 1 Foreword Iamhonoredtowritethisforewordfor‘TheNutanixDesignGuide’. Wehavealwaysbelievedthatcloudwillbemorethanjustarented modelforenterprises.Computingwithintheenterpriseisnuanced, asittriestobalancethefreedomandfriction-freeaccessofthe publiccloudwiththesecurityandcontroloftheprivatecloud.The privateclouditselfisspreadbetweencoredatacenters,remoteand branchoffices,andedgeoperations.Thetrifectaofthe3laws–(a) LawsoftheLand(dataandapplicationsovereignty),(b)Lawsof Physics(dataandmachinegravity),and(c)LawsofEconomics
    [Show full text]
  • Regarding the Challenges of Performance Analysis of Virtualized Systems Page 1 of 12
    Regarding the challenges of Performance Analysis of Virtualized Systems Page 1 of 12 Regarding the challenges of Performance Analysis of Virtualized Systems Steven Harris, [email protected] (A paper written under the guidance of Prof. Raj Jain) Download Abstract Virtualization is an irreplaceable technique for efficient management of hardware resources in the field of information technology and computer science. With this phenomenal technology come challenges that hinder proper performance analysis of virtualized systems. This study focuses on virtualization implementations, various enterprise solutions, characterization of observations, concerns, pitfalls, and industry standard virtualization performance measurement tools. Our goal is to delineate the various intricacies of virtualization to help researchers, information technologists, and hobbyists realize areas of concern and improvement in their performance analysis of virtualized systems. Keyword:Hardware Virtualization, Virtual Machine Monitor, Hypervisor, Performance Analysis, Benchmark, XEN, KVM, Wmware, Virtual Box, AMD-V, Intel VT, SPEC Table of Contents: • 1. Introduction ◦ 1.1 Virtualization Overview • 2. Virtualization Implementations ◦ 2.1 Full Virtualization ◦ 2.2 Partial Virtualization ◦ 2.3 Paravirtualization ◦ 2.4 Hardware Virtualization • 3. Virtualization Solutions • 4. Performance Analysis Challenges ◦ 4.1 Hypervisor Implementation ◦ 4.2 Guest Resource Interdependence ◦ 4.3 Guest Operating System ◦ 4.4 Utilization Ambiguity ◦ 4.5 Proper Model Construction • 5. Virtualization Measurement Tools ◦ 5.1 Benchmarks ◦ 5.2 SPECvirt_SC2010 ◦ 5.3 SPECweb2005 ◦ 5.4 SPECjAppServer2004 ◦ 5.5 SPECmail2008 ◦ 5.6 CPU2006 http://www.cse.wustl.edu/~jain/cse567-13/ftp/virtual/index.html 4/29/2013 Regarding the challenges of Performance Analysis of Virtualized Systems Page 2 of 12 ◦ 5.7 Network ◦ 5.8 Filesystems • 6. Conclusion • Acronyms • References 1.
    [Show full text]
  • Quantitative Comparison of Xen and KVM
    Quantitative Comparison of Xen and KVM Todd Deshane, Muli Ben-Yehuda Amit Shah Balaji Rao Zachary Shepherd, IBM Haifa Research Lab Qumranet National Institute of Jeanna N. Matthews Haifa University Campus B-15, Ambience Empyrean Technology Karnataka Mount Carmel, Haifa 31905 64/14 Empress County Surathkal 575014 India Computer Science Israel Behind Empress Gardens Clarkson University Pune 411001 India [email protected] Potsdam, NY 13699 USA [email protected] {deshantm, shephezj, amit.shah@ jnm}@clarkson.edu qumranet.com ABSTRACT For our initial set of tests, the experimental setup consisted of We present initial results from and quantitative analysis of two Ubuntu Linux 8.04 AMD64 on the base machine. The Linux leading open source hypervisors, Xen and KVM. This study kernel 2.6.24-18, Xen 3.2.1+2.6.24-18-xen, and KVM 62 were focuses on the overall performance, performance isolation, and all installed from Ubuntu packages. All guests were scalability of virtual machines running on these hypervisors. Our automatically created by a benchvm script that called debootstrap comparison was carried out using a benchmark suite that we and installed Ubuntu 8.04 AMD64. The guests were then started developed to make the results easily repeatable. Our goals are to with another benchvm script that passed the appropriate kernel understand how the different architectural decisions taken by (2.6.24-18-xen for Xen and 2.6.24-18 for KVM). The hardware different hypervisor developers affect the resulting hypervisors, system was a Dell OptiPlex 745 with a 2.4 GHz Intel Core 2 to help hypervisor developers realize areas of improvement for CPU 6600, 4 GB of RAM, 250 GB hard drive, and two 1 Gigabit their hypervisors, and to help users make informed decisions Ethernet cards.
    [Show full text]
  • World-Record Virtualization Performance: Cisco UCS B440 Server
    Cisco UCS B440 M1 High-Performance Blade Server: Performance Brief World-Record Virtualization Performance World-Record-Setting Blade Server Performance performance is observed. This procedure produces a VMware VMmark score and the The Cisco® UCS B440 M1 High-Performance Blade Server is a four-socket server number of tiles for the benchmark run. that leads the blade industry with a VMware VMmark benchmark score of 71.13 at 48 tiles, making it the highest-performing blade server for virtualiza- Industry-Leading Performance tion available anywhere. The server more than doubles HP’s closest blade Cisco tested the Cisco Unified Computing System equipped with a Cisco UCS B440 server result and delivers an 22 percent advantage over the closest M1 blade server containing four eight-core Intel Xeon X7500 series processors with 512 blade server score. This result means that customers can achieve higher consoli- GB of memory connected to two EMC CLARiiON CX4-480 storage systems through a dation ratios in their virtualized environments, with greater performance. Customers Fibre Channel SAN. This four-socket, 32-core system delivers a VMware VMmark score now can run even the most challenging workloads on a blade system with greater surpassing all blade server results posted at http://www.vmmark.com as of August 25, return on investment (ROI), reduced total cost of ownership (TCO), and the agility to 2010 (Figure 1). The VMware VMmark score of 71.13 is 243 percent higher than HP’s deploy applications more rapidly and securely. next-closest blade server result. It is 22 percent higher than the next-closest blade server result, the Dell PowerEdge M910 Blade Server with four eight-core Intel Xeon A Platform Built for Virtualization 7500 series processors.
    [Show full text]
  • Vmware Vsphere 4 Vs. Microsoft Hyper-V R2
    long gone are the days when “sun Microsystems” meant only Solaris on SPARC. peTeR BaeR gaLvin Sun is pushing hard to be a platform pro- vider for multiple operating systems, as well as a provider of their own Solaris operat- Pete’s all things Sun: ing system. In some ways this column is a VMware vSphere 4 vs. continuation of my April column (;login: Microsoft Hyper-V R2 April 2009, Volume 34, Number 2), which contained a virtualization guide. That col- Peter Baer Galvin is the chief technolo- umn discussed the virtualization offerings gist for Corporate Technologies, a premier created by Sun. This column explores the systems integrator and VAR (www.cptech. com). Before that, Peter was the systems rich, controversial, and important terrain manager for Brown University’s Computer Science Department. He has written articles of two of the market leaders in virtualiza- and columns for many publications and is tion, VMware and Microsoft. The topic is not co-author of the Operating Systems Concepts and Applied Operating Systems Concepts Sun-specific but is probably of interest to textbooks. As a consultant and trainer, Peter teaches tutorials and gives talks on security many Sun customers. The Sun x86 servers and system administration worldwide. Peter are certified to run VMware and Windows, blogs at http://www.galvin.info and twit- ters as “PeterGalvin.” as well as Solaris and Linux. In my experi- [email protected] ence, Sun x86 servers, especially the x4600 with its eight sockets and quad-core CPUs, make excellent virtualization servers. The question then becomes, which virtualiza- tion technology to run? OpenSolaris has Xen built in, but many sites want mainstream and well-tested solutions.
    [Show full text]
  • Vmmark 2.5 Virtualization Performance of the Dell Equallogic PS6210XS Storage Array
    VMMARK 2.5.2 VIRTUALIZATION PERFORMANCE OF THE DELL EQUALLOGIC PS6210XS STORAGE ARRAY Many modern data centers are using virtual machines (VMs) to consolidate physical servers to increase operational efficiency. As multi-core processors become more commonplace, underutilization of physical servers has become an increasing problem. Without virtualization, it is very difficult to fully utilize the power of a modern server. In a virtualized environment, a software layer lets users create multiple independent VMs on a single physical server, taking full advantage of the hardware resources. The storage solution, which is just as important as the servers and processors, should be flexible to accommodate the demands on real-world applications and operations that virtualization brings to the table. In all virtualized environments, storage performance can deteriorate due to a phenomenon called the input/output (I/O) blender effect. Multiple VMs send their I/O streams to a hypervisor for processing. Unfortunately, if you are using more than one type of workload, I/O profiles are no longer consistent. This randomization of I/O workload profiles, which can occur with all virtualization platforms, renders prior workload optimizations ineffective, which can increase latency, or response time. Because the performance requirements of storage in a completely virtualized environment differ from those in a physical or only partially virtualized environment, it is important to use a benchmark designed with these differences in mind, such as VMware VMmark 2.5.2. VMmark incorporates a variety of platform-level workloads NOVEMBER 2014 A PRINCIPLED TECHNOLOGIES TEST REPORT Commissioned by Dell such as vMotion® and Storage vMotion® in addition to executing diverse workloads on a collection of virtual machines.
    [Show full text]
  • TPC-V: a Benchmark for Evaluating the Performance of Database Applications in Virtual Environments Priya Sethuraman and H
    TPC-V: A Benchmark for Evaluating the Performance of Database Applications in Virtual Environments Priya Sethuraman and H. Reza Taheri, VMWare, Inc. {psethuraman, rtaheri}@vmware.com TPCTC 2010 Singapore © 2010 VMware Inc. All rights reserved Agenda/Topics Introduction to virtualization Existing benchmarks Genesis of TPC-V But what is TPC-E??? TPC-V design considerations Set architectures, variability, and elasticity Benchmark development status Answers to some common questions 2 TPC TC 2010 What is a Virtual Machine? A (VM) is a software computer that, like a physical computer, runs an operating system and applications. An operating system installed on a virtual machine is called a guest operating system. Virtual machines run on host servers. The same server can run many virtual machines. Every VM runs in an isolated environment. Started out with IBM VM in the 60s Also on Sun Solaris, HP Itanium, IBM Power/AIX, others A new wave started in the late 90s on X86 • Initially, enthusiasts ran Windows and Linux VMs on their PCs Traditional Architecture Virtual Architecture 3 TPC TC 2010 Why virtualize a server ? Server consolidation • The vast majority of server are grossly underutilized • Reduces both CapEx and OpEx Migration of VMs (both storage and CPU/memory) • Enables live load balancing • Facilitates maintenance High availability • Allows a small number of generic servers to back up all servers Fault tolerance • Lock-step execution of two VMs Cloud computing! Utility computing was finally enabled by • Ability to consolidate
    [Show full text]
  • Tuning IBM System X Servers for Performance
    Front cover Tuning IBM System x Servers for Performance Identify and eliminate performance bottlenecks in key subsystems Expert knowledge from inside the IBM performance labs Covers Windows, Linux, and VMware ESX David Watts Alexandre Chabrol Phillip Dundas Dustin Fredrickson Marius Kalmantas Mario Marroquin Rajeev Puri Jose Rodriguez Ruibal David Zheng ibm.com/redbooks International Technical Support Organization Tuning IBM System x Servers for Performance August 2009 SG24-5287-05 Note: Before using this information and the product it supports, read the information in “Notices” on page xvii. Sixth Edition (August 2009) This edition applies to IBM System x servers running Windows Server 2008, Windows Server 2003, Red Hat Enterprise Linux, SUSE Linux Enterprise Server, and VMware ESX. © Copyright International Business Machines Corporation 1998, 2000, 2002, 2004, 2007, 2009. All rights reserved. Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Contents Notices . xvii Trademarks . xviii Foreword . xxi Preface . xxiii The team who wrote this book . xxiv Become a published author . xxix Comments welcome. xxix Part 1. Introduction . 1 Chapter 1. Introduction to this book . 3 1.1 Operating an efficient server - four phases . 4 1.2 Performance tuning guidelines . 5 1.3 The System x Performance Lab . 5 1.4 IBM Center for Microsoft Technologies . 7 1.5 Linux Technology Center . 7 1.6 IBM Client Benchmark Centers . 8 1.7 Understanding the organization of this book . 10 Chapter 2. Understanding server types . 13 2.1 Server scalability . 14 2.2 Authentication services . 15 2.2.1 Windows Server 2008 Active Directory domain controllers .
    [Show full text]
  • Vmware® Vmmark™ Results
    VMware® VMmark™ V1.1.1 Results Vendor and Hardware Platform: Sun Microsystems Sun Fire [tm] X4270 VMmark V1.1.1 Score = Virtualization Platform: VMware ESX 4.0 (build 164009) 24.18 @ 17 Tiles Tested By: Sun Microsystems Test Date: 09/17/2009 Performance Section Configuration Section Notes Section Performance Configuration Notes for Workload Performance webserver javaserver mailserver fileserver database TILE_0 Actual Ratio Actual Ratio Actual Ratio Actual Ratio Actual Ratio GM p0 1699.12 1.67 18141.55 1.09 1630.25 1.49 16.81 1.31 2428.50 1.63 1.42 p1 1705.80 1.67 18133.20 1.09 1653.10 1.51 16.62 1.30 2442.20 1.64 1.42 p2 1657.72 1.63 18144.38 1.09 1531.58 1.40 16.89 1.32 2442.80 1.64 1.40 TILE_1 Actual Ratio Actual Ratio Actual Ratio Actual Ratio Actual Ratio GM p0 1714.58 1.68 17701.25 1.07 1538.97 1.40 16.77 1.31 2437.12 1.63 1.40 p1 1731.62 1.70 18148.62 1.09 1415.53 1.29 17.10 1.33 2447.72 1.64 1.39 p2 1734.97 1.70 18133.85 1.09 1370.42 1.25 17.19 1.34 2446.78 1.64 1.39 TILE_2 Actual Ratio Actual Ratio Actual Ratio Actual Ratio Actual Ratio GM p0 1801.25 1.77 18141.17 1.09 1709.15 1.56 17.43 1.36 2444.22 1.64 1.46 p1 1746.92 1.71 18132.33 1.09 1704.60 1.55 17.35 1.35 2455.28 1.65 1.45 p2 1720.03 1.69 18138.65 1.09 1522.97 1.39 17.53 1.37 2445.38 1.64 1.42 TILE_3 Actual Ratio Actual Ratio Actual Ratio Actual Ratio Actual Ratio GM p0 1602.20 1.57 18153.62 1.09 1596.67 1.46 16.67 1.30 2433.35 1.63 1.40 p1 1677.28 1.65 17684.88 1.06 1620.78 1.48 16.74 1.30 2442.82 1.64 1.41 p2 1686.78 1.66 18135.20 1.09 1569.42 1.43 16.61 1.29 2434.00
    [Show full text]
  • Vmware® Vmmark® Results for Cisco UCS B260 M4 Server
    VMware® VMmark® V2.5.1 Results Vendor and Hardware Platform: Cisco UCS B260 M4 Virtualization Platform: VMware ESXi 5.1.0 U2 Build VMmark V2.5.1 Score = 1483097 VMware vCenter Server : VMware vCenter Server 5.1.0 19.18 @ 16 Tiles Build 799731 Number of Hosts: 2 Uniform Hosts [yes/no]: yes Total sockets/cores/threads in test: 4/60/120 Tested By: Cisco Systems Test Date: 02-10-2014 Performance Section Configuration Section Notes Section Performance Configuration Notes for Workload Performance mailserver olio dvdstoreA dvdstoreB dvdstoreC TILE_0 Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS GM p0 328.10 0.99 124.00 4735.38 1.02 127.64 4076.12 1.85 62.41 2959.47 1.95 50.08 2249.15 2.13 39.09 1.51 p1 328.40 0.99 124.75 4712.48 1.02 133.04 4003.62 1.82 64.72 2921.20 1.92 51.58 2115.35 2.00 39.09 1.48 p2 326.07 0.99 134.00 4713.12 1.02 141.01 3948.88 1.80 66.91 3081.20 2.03 56.55 2281.50 2.16 43.61 1.51 TILE_1 Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS GM p0 326.40 0.99 131.75 4708.85 1.01 127.49 4112.73 1.87 60.72 3075.40 2.03 45.75 2159.25 2.04 37.11 1.51 p1 326.40 0.99 134.00 4692.15 1.01 133.32 3986.60 1.81 65.70 3035.95 2.00 52.75 2181.30 2.06 41.89 1.49 p2 325.98 0.99 134.00 4686.15 1.01 139.80 3901.30 1.77 69.07 2960.75 1.95 55.82 2204.45 2.08 46.73 1.48 TILE_2 Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS Actual Ratio QoS GM p0 328.75 1.00 124.00 4705.02 1.01 130.20 4016.22 1.83 64.63 2939.68 1.94 51.34 2090.45 1.98 40.01 1.48 p1 330.30
    [Show full text]
  • Multi-Dimensional Workload Analysis and Synthesis for Modern Storage Systems
    Multi-dimensional Workload Analysis and Synthesis for Modern Storage Systems A Dissertation Proposal Presented by Vasily Tarasov to The Graduate School in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Computer Science Stony Brook University Technical Report FSL-13-02 April 2013 Abstract of the Dissertation Proposal Multi-dimensional Workload Analysis and Synthesis for Modern Storage Systems by Vasily Tarasov Doctor of Philosophy in Computer Science Stony Brook University 2013 Modern computer systems produce and process an overwhelming amount of data at a high and growing rates. The performance of storage hardware components, however, cannot keep up with the required speed at a practical cost. To mitigate this discrepancy, storage vendors incorporate many workload-driven optimizations in their products. New emerging applications cause workload patterns to change rapidly and significantly. One of the prominent examples is a rapid shift towards virtualized environments that mixes I/O streams from different applications and perturbs their access patterns. In addition, modern users demand more and more convenience features: deduplication, snapshotting, encryption, and other features have become almost must-have features in modern storage solutions. Stringent performance requirements, changing I/O patterns, and the ever growing feature list increase the complexity of modern storage systems. The complexity of design, in turn, makes the evaluation of the storage systems a difficult task. To resolve this task timely and efficiently, prac- tical and intelligent evaluation tools and techniques are needed. This thesis explores the complex- ity of evaluating storage systems and proposes a Multi-Dimensional Histogram (MDH) workload analysis as a basis for designing a variety of evaluation tools.
    [Show full text]
  • An Extensible I/O Performance Analysis Framework for Distributed Environments
    An Extensible I/O Performance Analysis Framework for Distributed Environments Benjamin Eckart1, Xubin He1,HongOng2, and Stephen L. Scott2 1 Tennessee Technological University, Cookeville, TN 2 Oak Ridge National Laboratory, Oak Ridge, TN Abstract. As distributed systems increase in both popularity and scale, it becomes increasingly important to understand as well as to system- atically identify performance anomalies and potential opportunities for optimization. However, large scale distributed systems are often complex and non-deterministic due to hardware and software heterogeneity and configurable runtime options that may boost or diminish performance. It is therefore important to be able to disseminate and present the informa- tion gleaned from a local system under a common evaluation methodol- ogy so that such efforts can be valuable in one environment and provide general guidelines for other environments. Evaluation methodologies can conveniently be encapsulated inside of a common analysis framework that serves as an outer layer upon which appropriate experimental de- sign and relevant workloads (benchmarking and profiling applications) can be supported. In this paper we present ExPerT, an Extensible Performance T oolkit. ExPerT defines a flexible framework from which a set of benchmarking, tracing, and profiling applications can be correlated together in a unified interface. The framework consists primarily of two parts: an extensible module for profiling and benchmarking support, and a unified data discov- ery tool for information gathering and parsing. We include a case study of disk I/O performance in virtualized distributed environments which demonstrates the flexibility of our framework for selecting benchmark suite, creating experimental design, and performing analysis. Keywords: I/O profiling, Disk I/O, Distributed Systems, Virtualization.
    [Show full text]